Kopite7kimi, a widely known leaker on Twitter, has printed what will be the first specs of Nvidia’s upcoming GeForce RTX 40-series graphics playing cards household. Since we’re a number of months away from Nvidia’s launch of its codenamed Ada Lovelace structure, it is potential there shall be adjustments, however given the leaker’s observe report, the specs are no less than value contemplating.
Nvidia’s preliminary GeForce RTX 40-series household will include three graphics playing cards: the GeForce RTX 4090, based mostly on the AD102-300 graphics processor paired with 24GB of GDDR6X reminiscence; the GeForce RTX 4080, powered by the AD103-300 GPU mated with 16GB of GDDR6/GDDR6X reminiscence; and the GeForce RTX 4070 that includes the AD104-275 chip geared up with 10GB of GDDR6.
The variety of CUDA cores featured by Nvidia’s Ada Lovelace GPUs present that the brand new graphics boards will doubtless have a substantial efficiency improve when in comparison with Nvidia’s current GeForce RTX 30-series, that are amongst greatest graphics playing cards round.
Rumored Nvidia GeForce RTX 40-Sequence Specs
| GPU | FP32 CUDA Cores | Reminiscence Configuration | TBP | |
| GeForce RTX 4090 | AD102-300 | 16384 | 24GB 384-bit 21GT/s GDDR6X | 450W |
| GeForce RTX 4080 | AD103-300 | 10240 | 16GB 256-bit 18GT/s GDDR6? | 420W |
| GeForce RTX 4070 | AD104-275 | 7168 | 10GB 160-bit 18GT/s GDDR5 | 300W |
All these preliminary GeForce RTX 40-series graphics boards will allegedly be fairly energy hungry with the top-of-the-range product that includes a thermal board energy (TBP) of 450W and different rated for 420W and 300W, respectively. In relation to energy consumption, the data corroborates with earlier rumors indicating a significant improve of TBP for Ada Lovelace household.
Whereas energy consumption of Nvidia’s next-generation GeForce RTX 40-series merchandise will improve in comparison with current-generation graphics playing cards, not every part is obvious about pricing of the brand new boards. Kopite7kimi tells us to not anticipate GeForce RTX 4000 boards to function decrease MSRPs. Worth is a very murky matter, maintaining in thoughts larger manufacturing prices on TSMC’s N5 course of expertise (or quite personalized 4N) when in comparison with Samsung’s 8LPP.
It’s noteworthy that this time round Nvidia will use three completely different graphics processors for 3 preliminary playing cards. This may be a results of first rate preliminary yields of Nvidia’s Ada Lovelace GPUs at TSMC, or maybe because of some adjustments within the firm’s inner chip design schedules. Anyhow, the gaps between the variety of CUDA cores featured by the AD102, AD103 and AD104 GPUs would possibly point out that Nvidia shall be very versatile with configurations for refresh (Ti-branded or Tremendous-branded merchandise).
In fact, Nvidia usually doesn’t touch upon specs of unreleased merchandise, so take the data with a grain of salt.
